INDEX

Explanations

words related to the term "engagement" or variations thereof

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "eng"

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ô

-0.69

pmwiki

-0.65

 mileage

-0.65

phis

-0.65

IFIED

-0.64

cule

-0.63

staking

-0.62

*/(

-0.62

ittee

-0.59

istically

-0.58

POSITIVE LOGITS

eng

1.29

edin

0.95

iment

0.93

raved

0.91

ravings

0.88

eers

0.88

rave

0.85

ugu

0.85

uru

0.84

rador

0.83

Activations Density 0.005%

words related to the term "engagement" or variations thereof

occurrences of the substring "eng"

No Comments

No Known Activations